After the first four years of our Bioengineering Research Partnership, implantable devices for epilepsy are now a reality. This is due, in part, to translation of technology developed by our group to industry. Data from multi-center clinical trials of first generation responsive antiepileptic devices indicate that this new technology is safe, and that there is promise of significant benefit to patients. They also demonstrate that 1st-generation devices rarely make patients seizure free. This is because we do not yet understand when, where and how to deliver electrical stimulation to pre-empt seizures, or the mechanisms underlying seizure generation in epileptic networks. These challenges, and translating them into more effective second-generation devices, are the focus of this proposal. Specifically, our aims are: (1) To understand mechanisms underlying seizure generation in two well characterized, spontaneously seizing animal models of epilepsy with documented similarities to refractory human epilepsy, (2) To map seizure generation in the epileptic network to determine where to place sensing electrodes and when to stimulate to maximize seizure suppression and minimize side effects. (3) To develop more effective closed loop stimulation algorithms for controlling seizures. These Aims will be accomplished through a series of projects led by established collaborators in neurology, neuroscience, bioengineering and industry, at Penn, CHOP, Georgia Tech, and BioQuantix, Inc. Teams will focus on improving upon results from first-generation human devices through detailed animal experiments on multiple temporal and spatial scales. These include: (1) the cellular level, through broad-band unit recording and biophysically accurate computational modeling;(2) the network level, with in vitro experiments on hippocampal slices using voltage sensitive dyes and multi-electrode arrays;and (3) the whole brain level, through simultaneous micro and macroelectrode field recordings and responsive brain stimulation in vivo. These experiments will build upon the substantial progress made during the first cycle of our Bioengineering Research Partnership grant. The unique composition of our group, its track record of successful technology transfer, and our ability to learn from and immediately convey our discoveries to existing programmable devices, provide an unprecedented opportunity to perform cutting-edge neuroscience and bioengineering research and immediately translate it into better treatment for patients.